Preventing Hysteresis in homemade Force Sensing Resistors

[JohnnySmithy]

Hello, I am new to the forums and physics in general. I hope I am not posting in the wrong place. My understanding of physics in general is rudimentary at best, so please let me know if any of my statements are incorrect.

I understand this is an extremely long post, so feel free to jump to "The Problem" below if you are already familiar with what I'm talking about.

Background:

My friends and I have started a new hobby project of trying to create a wearable Force Sensing Resistor (FSR) to take rough readings of forces around the ranges of 1-100 pounds. The wish is that we would be able to create a very dumb-downed version of this device. We thought it would be very cool to be able to pick something up and have the device let us know how much weight we are carrying.

We decided to focus on several materials whose resistance changes inversely when a force is applied on it. We were hoping to map out the resistance output of the material by applying known weights on it. Essentially, we were hoping to regress the predetermined weights (the Y-variable) on the resistance output (the X-variable), so that when an unknown object on the sensor produces a resistance output (X), we can predict the weight of it (Y).

The Materials and Set Up
We decided to focus on "smart" materials, such as velostat and quantum tunneling composite (aka QTC). We didn't focus on the traditional conductive ink printed FSRs because the ones available online seemed a bit bulky/rigid for our applications, though we should definitely explore it.

1. Velostat
   
   Velostat is a type of conductive material that can be commonly found in packaging materials for electronics to prevent static​
   With velostat, we were hoping to recreate this device on page 2: WARNING .PDF FILE​
2. Quantum Tunneling Composite (QTC)
   
   QTC is a strange type of material that is an inductor at rest and a conductor when stressed​
   With QTC, we were hoping to recreate something similar to the device on page 2, application 3: https://www.mindsetsonline.co.uk/images/QTC.pdf​

Our setup was pretty basic: the sensor was attached to the multi-meter with the meter set to measure resistance. We laid something flat on top of the sensor so that any weight on top of the sensor would be distributed evenly. We then applied the weight and measured the resistance.

The Problem
To cut to the chase, the resistance output measured by the multi-meter was never stable for any of the materials. We couldn't establish a relationship between weight/force and resistance output since the resistance would be decaying at a (seemingly) unpredictable rate. We surmise that the material is adapting to the weight and forming around it, lowering the resistance.

How can we solve this "hysteresis"? Is there any way to work around it? Any advice on any aspect of our project would be appreciated.

Also, we've been trying to get our hands on a type of material called Eeoynx, which seemingly produces consistent readings, as show in this Youtube video. The only problem is that this material is incredibly expensive and hard to get. Does anyone have any experience with this type of material?

Anyway, I hope my post didn't go too long. I appreciate any help or advice. Thank you.

 

[0xDEADBEEF]

You could try using a strain gauge. I think most force sensors suffer from drift effects so often you just measure changes of force.

 

[JohnnySmithy]

How big are most strain gauges? Can I realistically "shrink" it so that I can use multiple strain gauges to cover areas of the hand?

Also, do I need to apply it to a completely flat surface? How much of an effect does the non-flat surface of the hand have on the gauge?

Thank you.

 

[0xDEADBEEF]

Well I guess with a glove it wouldn't work too well. The nice thing about the strain gauge is, that is measures geometry and thus the underlying material defines the proportionality constants, but I don't think it would produce a lot of useful signal if you bend it in you hand by gripping something.